Deepfake: ITV's Mary Nightingale on how she was 'livid' following AI manipulated video

ITV News journalist and presenter Mary Nightingale has spoken of how outraged she was to discover she had been a victim of a deep fake last month.

A manipulated version of the news anchor appeared on social media in February before it was deleted.

The video, which lasted around 45 seconds, showed Nightingale sitting at her desk in an ITV studio as if presenting the ITV Evening News. 

However, it then went on to promote a type of investment app in which singer Dua Lipa was also deepfaked.

The video was deleted shortly after it first appeared online.

Speaking to ITV Cymru’s Wales This Week programme, Nightingale said: “It made me absolutely livid.

“For someone to take my image and supposedly my voice and manipulate it in that way, it’s theft, it’s identity theft.”

Deepfake videos and pictures have become more common as generative Artificial Intelligence (AI) tools become more widespread and accessible. 

As many as 500,000 appeared online last year, and fake celebrity endorsements are also becoming more common.

In the UK, news broadcasters adhere to strict rules that limit their ability to promote products, espeicially while on air.

With elections in both the UK and USA later this year, there are also fears they could be used to spread disinformation.

“I looked at it with my husband and he said ‘okay I know it’s not you’ but anyone who didn’t know it wasn’t me would actually find it quite convincing,” Nightingale continued.

“In our job, trust is everything. If people don’t trust you, you cannot effectively do your job. 

“If people start to doubt my credibility, if people start to think ‘oh she would say that, she flogs dodgy investment deals, then that really does start to eat away at my professional credibility.”

The developments of AI technologies, especially those known as machine learning - where machines learn from their own actions - have caused many people to worry about the AI becoming too powerful.

Advancements in generative AI technology have made tools once limited to highly skilled professionals accessible and cheap. Credit: PA

However, Professor Matt Jones from Swansea University’s department of computer science believes it’s deepfakes and sinister uses that present the most tangible and immediate dangers.

“Deepfakes is one of the problems that we should address,” he said.

“There’s a great deal of worry about AI becoming our overlords, becoming sentient, but I actually think that’s a smokescreen.

“What we need to worry about is people using things like deepfakes to manipulate elections or to scam you out of your money. 

“We should focus on those near and present dangers and not get carried away worrying about Sci-Fi visions of sentient AI.”

Trade unions and campaigners have called for stronger legislation and regulation on AI and its uses. 

The House of Lords is currently debating the introduction of an ‘AI Authority’, and an Artificial Intelligence (Regulation) Bill that would assess and monitor potential risks to the economy, security, accountability and transparency.

Watch Wales This Week: Ready for AI Revolution on Tuesday at 8pm on ITV 1 Cymru Wales and catch up afterwards on ITVX.